Session 7

Back to Schedule

Title of session: Effective partnerships for quality improvements

Chair: Claudia Junker

Room: S3A Barbakan

Time: 11:30 - 13:00

Date: 28 June

Session 7 - papers & presentations

Presenting AuthorAbstract
Yolanda Gómez Menchón
Title: <<< Legal aspects related to enhance the cooperation with data holders: The case of Spain. >>>
The concept of cooperation is closely related to coordination in the Regulation 223/2009 and in the Code of Practice of European Statistics. Cooperation is understood as an horizontal tool between national statistical authorities and other public entities or between statistical authorities and the scientific community, or even among statistical authorities of different countries (international cooperation), but the ESS Quality Declaration adopted in 2016 introduces a new meaning of cooperation. The first question that comes to mind is if the ESS is moving from a traditional cooperation among public data providers -the owners of the administrative data- to an enlarged cooperation adding the private sector. Addressing a very hot issue like big-data, in this article we analyse some issues such as the need for legal support for cooperation with data holders or which type of legal acts could be more useful for the statistical community.
Albrecht Wirthmann
Title: <<< Ethical implications of using Big Data for Official Statistics >>>
We are currently experiencing an all-embracing digitalisation of our societies and economies. The pervasive nature of information and communication technologies is leading to the 'datafication' of most of our activities and relationships. This development is producing so far unknown amounts of data. New developments in Information Technologies do not only allow capturing but also storing, linking and analysing these data to infer conclusions on data subjects in massive amounts. This new ability might impact persons' everyday life in different ways, positively or negatively. Ethical implications thereof have already been discussed in literature. The proposed paper will discuss the implications of using big data in official statistics. These implications will differ from risks and hazards related to big data usage in general. Firstly, Statistical offices are not targeting individual subjects but only aggregated results and identified patterns. Secondly, there are already principles and ethical guidelines concerning the statistical confidentiality of personal data. Therefore the paper will focus on issues of data quality, availability of data sources, dependence on third party sources, data manipulation influencing the results, transparency of data and methods of data analytics, or scientific approach to data analytics. The paper will identify issues and analyse possible consequences for official statistics and contrast them with the existing ethical frameworks (Fundamental principles of official statistics, European Statistics code of practice, Declaration on Professional Ethics). Finally the paper will define or emphasize relevant principles on the ethics of using big data that could be followed in the process of integrating these new data sources into Official Statistics.
Hélène Bérard
Title: <<< Innovation in collection: strategies to improve and maintain the quality of administrative data >>>
Statistics Canada actively obtains administrative data from the public sector, the private sector and various organizations to support the production of official statistics. A number of strategies are in place to facilitate data access and to ensure that the data obtained from various organizations meet specific quality requirements. The strategies implemented to meet these goals include: 1) outreach activities prior to obtaining the data to explain the usefulness of the data, 2) offering technical support, 3) organizing data quality workshops to promote the use of sound methods, 4) identifying specific requirements in data acquisition agreements to ensure the quality and timeliness of the data, 5) negotiating access to a test file to assess fitness for use before proceeding with an official acquisition, 6) establishing a Quality Evaluation Framework to ensure consistency and completeness of assessments across the departments many administrative data files and, 7) maintaining post acquisitions communications, and 8) developing mutually beneficial relationships to ensure long term supply, usefulness and quality of the administrative data. One of our current challenges is to identify and influence future changes to administrative data sources that may affect statistical use. This paper will discuss specific examples to illustrate the challenges faced and the lessons learned in establishing these various strategies.
Natalie Rosenski
Title: <<< Access to Big Data for statistical purposes >>>
The topic big data is of high importance for all statistical offices as well as the Federal Statistical Office of Germany (Destatis). On this account, Destatis is a member of the European Statistical System (ESS) Task Force, the ESS Steering Group as well as the ESSnet Big Data. Furthermore, Destatis is a member of the United Nations Global Working Group on Big Data for Official Statistics. Destatis has already started several feasibility studies concerning big data, e.g. web scraping for price and labour market statistics. Furthermore, Destatis has entered cooperations with T-Systems, which is a subsidiary of the Deutsche Telekom AG, and the German Aerospace Center to use mobile phone and satellite data for the examination of its usability in official statistics. To be able to use big data for official statistics and not only for feasibility studies, individual cooperations with private enterprises are not sufficient. Instead, a permanent access to data held by private enterprises is necessary, which is unattached of their willing or the market situation. Therefore, Destatis is aiming for a legal basis to get access to privately held data, on which further cooperations can be premised on. Big data should then be free of charge, but the service of the enterprises to prepare the data for statistical purposes could still be charged by the enterprises. Fees for this service enable the access to the knowledge of the data providers. Furthermore, official statistics could get semi-final products by the data providers in order to save money within the offices. A benchmark for this legal foundation is delivered by France and the United Kingdom.
Peter Struijs
Title: <<< Big Data Strategies for Official Statistics >>>
The environment of National Statistical Institutes (NSIs ) is rapidly changing in many respects. The emergence of new data sources provides a number of opportunities for official statistics. At the same time this creates challenges, including how to deal with quality issues when for example developing so-called experimental statistics and turning them into official statistics. As many NSIs have started using big data for statistics, the need for a strategic approach has become increasingly clear. The paper describes and assesses strategic options and explains the big data strategy of Statistics Netherlands, which is now being implemented in the Dutch Center for Big Data Statistics. In essence, big data strategies are about positioning NSIs in the changing environment. The paper identifies the true game-changers for official statistics and formulates the associated strategic questions. Their answers depend on where the value added of official statistics is sought, which is to no small extent related to quality considerations associated with the use of new data sources. New approaches may be called for. In any event, the role of NSIs is bound to change. The traditional role of quasi monopolistic provider of statistics on the many facets of society will erode through the rise of competition. However, the institutional and professional foundation of NSIs may also be exploited for assuming new roles. Ideally this will result in a society that is better informed about relevant phenomena and better equipped to counter tendencies where the value of facts is discredited.

Back to Schedule

Font Resize